Tier 1 to 4: The four tier classes of the data center

Imagine your company losing millions of euros within a few hours due to a data center outage. Or your customers and employees are unable to access your services for several days. The Tier system was developed to avoid such scenarios and make data centers comparable.

The tier topology for classifying data centers was introduced as a global standard at the end of the 1990s by the Uptime Institute based in the USA. Each "tier" (level) stands for a certain rank that the respective data center or its subsystems fulfill. It is the most frequently used "standard" to describe the structure and availability of a data center.

The tier topology provides for clearly defined availability levels (Tier 1 to Tier 4). These levels are based on measurable criteria for redundancy, maintenance concepts and actual operating times. Tier 1 is the least reliable environment, while Tier 4 is classified as "highly available". Choosing the right availability is crucial for the success or failure of digital business models. Although many providers advertise Tier 4, practice shows that Tier 3 is the more economically viable solution in most cases.

Availability as a key criterion

Availability refers to the probability that a system will function as planned at a given time. It is calculated as the ratio of downtime to total time:

Availability (%) = 1 - downtime / (production time + downtime)

What does "high availability" mean?

For High availability the probability that a system is available must be above 99.99%, which means that the annual downtime is in the range of minutes.

The four availability levels of the data center

Tier 1: Basic infrastructure

• Availability: 99.67% (28.8 hours of downtime per year)
• Characteristic: No redundancy available
• Features:
  • Single supply route for energy and cooling
  • Maintenance only possible with complete interruption of operation
  • High risk of failure due to single points of failure
  • Lowest energy density/cooling load: 220-320 watts/m²

Area of application: Small companies with non-critical workloads

The Tier 1 standard describes a basic infrastructure structure that does not include redundancy and is therefore only suitable for non-critical workloads (loads). Features are simply existing components for supplying and distributing the loads. In this setup model, maintenance is only planned and only possible if the systems are completely interrupted. There is also a high risk of failure, as faults can occur at all connection and supply levels due to technical components and human error.

Tier 2: Simple redundancy

• Availability: 99.75% (22 hours of downtime per year)
• Characteristic: N+1 redundancy for critical components
• Features:
  • Redundant UPS modules, cooling units and generators
  • Still only one distribution path
  • Planned and unplanned events can lead to downtime
  • Energy density/cooling load: 430-540 watts/m²

Area of application: Small and medium-sized enterprises, in-house data centers, data backups and non-critical databases

In a Tier 2 data center, redundancy is provided by redundant power and cooling components ("N+1" redundancy), such as an additional UPS module, additional cooling units, pumps or generators. However, the distribution path for all loads remains simple, which is why availability is increased but is significantly lower than the following Tier 3 level. In this setup, which is possibly intended as a transitional solution, planned measures and unplanned events can still severely impair the running systems or lead to a complete failure.

Tier 3: The market standard - maintenance-tolerant systems

• Availability: 99.98% (1.6 hours of downtime per year)
• Characteristic: Maintenance possible during operation
• Features:
  • N+1 redundancy for all critical systems
  • Multiple, active and passive supply routes
  • Continuous maintenance without interrupting operations
  • Several fire compartments for increased safety
  • Energy density/cooling load: 1,070-1,620 watts/m²

Area of application: Companies with high availability requirements

• E-commerce and online trading (continuous sales processing required)
• SaaS providers and cloud services (Service Level Agreements with customers)
• Fintech and digital financial service providers (time-critical transactions)
• Large companies with business-critical IT systems and customer databases
• Telecommunications provider and IT service provider
• Production company with IT-controlled production processes

Tier 3 is the most widely used standard. A significant difference to Tier 2 is that "continuous maintenance" is possible in a Tier 3 data center. Maintenance of individual components can be planned and carried out without interrupting ongoing operations. There is "N+1" redundancy for the supply and distribution of the load (please also read our blog article "How does redundant power work in the data center?"). The maximum downtime (average over 5 years) is thus reduced to 1.6h per year.

Tier 4: Fault-tolerant infrastructure

• Availability: 99.995% (26 minutes of downtime per year)
• Characteristic: Complete fault tolerance
• Features:
  • Complete redundancy of all systems (2N redundancy)
  • No single points of failure
  • Automatic reaction and defense mechanisms
  • Physically isolated, complementary systems
  • Energy density/cooling load: over 1,620 watts/m²

Area of application: Mission-critical applications with absolute availability requirements

• Critical infrastructure (KRITIS)Energy suppliers, water supply, traffic control centers
• FinanceStock exchanges, central banks, major banks with systemically important functions
• HealthcareUniversity hospitals, emergency centers with life support systems
• Government and authoritiesSecurity authorities, civil protection, military
• AerospaceAir traffic control, space control centers
• ResearchScientific data centers of international importance

Fault-tolerant infrastructure elements are added to the current maximum level Tier 4. In this setup, each system element has its own automatic mechanisms that start automatic reaction and defense mechanisms to counteract any errors or failures of individual components. All capacity and supply elements are also redundant throughout and designed to be technically optimally coordinated with each other. Complementary systems and supply paths are physically isolated from each other in order to avoid the risk of "contagion" in the event of failures. A Tier 4 data center is considered "highly available" with an average availability of 99.995% per year, but is very complex to set up and correspondingly expensive to implement.

Many modern data centers implement Tier 3+ configurations that integrate individual Tier 4 elements without taking on their full complexity. These hybrid approaches offer enhanced security at an economically reasonable cost.

Fault tolerant systems

Fault-tolerant systems achieve particularly high availability because they can react to almost all conceivable causes of error with the help of intelligent software. In addition, the design of fault-tolerant systems eliminates causes for Single Points of Failure (SPOF). An SPOF refers to an individual component that is essential for the correct and reliable functioning of the overall system. This also includes the design of the network and storage technology: for example, a failed network switch can mean that the service of the entire network is no longer available.

Through the production of Hardware and networkRedundancy and automatic load balancing, SPOF risks can be mitigated. To achieve this, the individual hardware and network components such as routers and switches of the same type are created multiple times. In the event of a failure, the redundant component can take over the task of the others. Standby systems (mirrored computer hardware) and  Automatic failover mechanisms ensure immediate switchover in the event of component failure.

Organizational aspects of high availability

High availability is not only achieved through technical infrastructure. The organizational and executive structures are just as important for the secure operation of the infrastructure. These include, for example

• trained service personnel with 24/7 availability
• Strategic provision of critical components and Spare parts
• Comprehensive Maintenance contracts with guaranteed response times
• Documented emergency procedures for different scenarios
• Precise communication structures for fault messages
• Event logging for tracking and optimization

In the world of data centers, other certification standards are relevant in addition to tier classification:

• ISO/IEC 27001: Information Security Management Systems (ISMS)
• ISO/IEC 20000: IT Service Management (ITSM)
• Basic IT protection
• EN 50600: European standard for data center planning and operation
• TSI (Trusted Site Infrastructure): German alternative to animal classification

For the ISO/IEC 27001 and ISO/IEC 20000 standards, there are also supplementary guidelines with best practice instructions. This refers to ISO/IEC 27002 and ITIL (IT Infrastructure Library).

Conclusion on the right choice of animal

Tier classification is a tried and tested framework for availability requirements. However, reality shows clear preferences. Tier 3 represents the current practical standard for modern data centers. With 99.98% availability and easy-to-maintain systems, Tier 3 meets virtually all requirements for business-critical applications. The ability to perform maintenance without interrupting operations makes this standard particularly valuable for companies with continuous service requirements. True Tier 4 data centers are extremely rare in Germany and only justify themselves for absolutely mission-critical applications such as critical infrastructure, system-relevant financial institutions or government institutions. The jump from Tier 3 to Tier 4 often doubles the costs with a marginal gain in availability.

Recommended links:

Wolfgang Heinhaus, Ulrike Ostler (2017, June 30) What secures data centers from Tier I to IV? URL: https://www.datacenter-insider.de/was-sichert-rechenzentren-von-tier-i-bis-iv-a-619349/

Andrea Held (2015, October 6) High Availability and Downtime: An Introduction URL:https://www.informatik-aktuell.de/betrieb/verfuegbarkeit/hochverfuegbarkeit-und-downtime-eine-einfuehrung.html

Holger Skurk (2013, December 13) "Betriebssicheres Rechenzentrum" URL: https://www.bitkom.org/Bitkom/Publikationen/Betriebssicheres-Rechenzentrum.html

Uptime Institute (2018, January) Tier Standard: Topology, Uptime Institute URL: https://uptimeinstitute.com/uptime_assets/d60e4b92ffa912e586a68d76617c4f12c3700681477064382a03207927a1ee96-00001A.pdf